Release v1.0.0
We are excited to announce the first official release of the Hungarian Network (Hnet) repository! This initial version lays the foundation for multi-source localization (DoA estimation) using the deep learning-based Hungarian algorithm implementation. Below are the key features, functionalities, and components included in this release.
🔑 Key Features
1. Synthetic Data Generation
generate_hnet_training_data.py- Generates synthetic distance matrices and association matrices essential for training the Hnet model.
- Supports various configurations for angular resolutions and DoA combinations.
2. Model Training
run.py- Main script to train the Hnet model using the generated datasets.
- Utilizes PyTorch Lightning for streamlined training loops and Hydra for configuration management.
3. Configuration Management
configs/Directory- Contains YAML configuration files.
- Facilitates easy customization of training parameters, logging settings, and more.
4. Testing and Coverage
test.py- Comprehensive test suite to validate model functionality and performance.
pytest.ini- Configuration file for Pytest to manage test settings.
tests/Directory : consistency and scenario tests are located intests/directory.
htmlcov/Directory- Contains HTML coverage reports to monitor test coverage and identify untested code segments.
6. Continuous Integration (CI)
/.github/workflows/ci-cd.yml- GitHub Actions workflow file to automate testing, coverage checks, and Docker image builds upon commits and pull requests.